CN114206209A

CN114206209A - Orthostatic hypotension detection system and method using heart rate based machine learning algorithm and wearable measurement device

Info

Publication number: CN114206209A
Application number: CN202080056181.3A
Authority: CN
Inventors: 金柄助; 金正彬
Original assignee: Tubei Dtx Co ltd
Current assignee: Tubei Dtx Co ltd
Priority date: 2019-08-07
Filing date: 2020-08-06
Publication date: 2022-03-18
Also published as: US20220280046A1; WO2021025507A1; KR102290274B1; KR20210017270A

Abstract

An orthostatic hypotension detection system using a heart rate based machine learning algorithm according to an embodiment of the invention may include: an input part for receiving variables including at least one of an E-I difference and an E: I ratio calculated from the age, blood pressure and heart rate of a patient and a Wal-Save ratio calculated from the Wal-Save movement; and a determination section that determines whether the patient has orthostatic hypotension according to a learned machine learning algorithm based on the variable received through the input section.

Description

Orthostatic hypotension detection system and method using heart rate based machine learning algorithm and wearable measurement device

Technical Field

The invention relates to an orthostatic hypotension detection system and method using a heart rate based machine learning algorithm and a wearable measuring device.

Background

Orthostatic Intolerance (OI) refers to autonomic nervous system dysfunction in which the patient stands up, lying or sitting, with reduced blood flow to the brain and heart, resulting in light vertigo, dizziness, blurred vision, palpitations, nausea, and fatigue.

Orthostatic Hypotension (OH), defined as the lowering of systolic pressure by more than 20mmHg and diastolic pressure by more than 10mmHg in standing up, does not produce an increase in heart rate to maintain blood flow, is one of OI. OH can occur in a variety of diseases associated with the autonomic nervous system, such as parkinson's disease, multiple atrophy, pure autonomic failure and diabetic autonomic neuropathy, and is associated with increased risk of fall injury, cardiovascular events and cognitive disorders. Therefore, it is necessary to detect OH early and manage it in a timely manner.

For the diagnosis of OH, the vertical Tilt Test (HUT; Head-Up Tilt Table Test) is widely used. However, many patients have reasons for contraindication of HUT, e.g. failure to maintain posture on a tilting table due to physical disability, severe anemia, kidney or heart failure, heart valve disease, severe coronary artery disease, and acute and subacute cerebral stroke or myocardial infarction. In addition to the physical limitations described above, repeated HUT to monitor OH treatment response and symptom progression has limitations due to the time and cost burden required to perform HUT, and the limitation of high probability of false negative due to OH generated by various stimuli in daily life being performed only in a set laboratory environment.

In order to overcome the limitations of HUT as described above, research for finding biomarkers for alternative schemes for diagnosing OH is being conducted, but there is still a lack of a technique for timely and accurate diagnosis of OH using heart rate variation that can be acquired by non-postural stimulation (non-postural stimuli) without HUT in daily life.

In order to solve the above problem, an embodiment of the present invention provides an orthostatic hypotension detection system using a machine learning algorithm based on a heart rate measured in a non-postural stimulus.

The orthostatic hypotension detection system using a heart rate based machine learning algorithm may include: an input part for receiving variables including at least one of an E-I difference and an E: I ratio calculated from the age, blood pressure and heart rate of a patient and a Wal-Save ratio calculated from the Wal-Save movement; and a determination section that determines whether the patient has orthostatic hypotension according to a learned machine learning algorithm based on the variable received through the input section.

Further, the orthostatic hypotension detection system using a heart rate based machine learning algorithm may include: a wearable measurement device worn on a patient and measuring a heart rate of the patient; a processing device for calculating an E-I difference and an E: I ratio from the heart rate measured by the wearable measuring device, and calculating a valsalva ratio from the heart rate measured by the wearable measuring device according to a valsalva maneuver; and a determination unit for determining whether the patient has orthostatic hypotension based on the E-I difference, the E: I ratio, and the Walsajou ratio calculated by the processing unit, according to a learned machine learning algorithm.

In another aspect, a method of orthostatic hypotension detection using a heart rate based machine learning algorithm is provided according to another embodiment of the invention.

The orthostatic hypotension detection method using a heart rate-based machine learning algorithm may include the steps of: calculating an E-I difference and an E: I ratio from the heart rate of the patient; calculating a valsalva ratio according to the valsalva maneuver; and determining whether or not orthostatic hypotension is present according to a machine learning algorithm based on the E-I difference, the E: I ratio, and the Walsawa ratio.

In another aspect, a wearable measurement device is provided in accordance with another embodiment of the present invention.

The wearable measurement device worn on a patient's body and measuring the patient's heart rate is loaded with software implementing an orthostatic hypotension detection method using a machine learning algorithm based on heart rate, and the software determines whether the patient has orthostatic hypotension based on the measured heart rate.

The means for solving the above problems do not list all the features of the present invention. Various features of the present invention, together with advantages and effects thereof, may be understood in more detail with reference to the following detailed description.

According to an embodiment of the present invention, it is possible to accurately diagnose OH in time using heart rate changes that can be acquired by non-postural stimulation without HUT in daily life.

Drawings

Fig. 1 is a block diagram of an orthostatic hypotension detection system using a heart rate based machine learning algorithm according to an embodiment of the present invention.

Fig. 2 is a block diagram of an orthostatic hypotension detection system using a heart rate based machine learning algorithm according to another embodiment of the present invention.

Fig. 3 is a flow chart of an orthostatic hypotension detection method using a heart rate based machine learning algorithm according to another embodiment of the present invention.

Detailed Description

Hereinafter, preferred embodiments will be described in detail with reference to the accompanying drawings so that those having ordinary knowledge in the art to which the present invention pertains can easily carry out the present invention. However, in describing the preferred embodiments of the present invention in detail, if it is judged that the related well-known functions or detailed description of the configurations may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. Further, the same reference numerals are used throughout the drawings for portions having similar functions and actions.

In addition, when one portion is "connected" to another portion throughout the specification, the case of "directly connecting" is included, and the case of "indirectly connecting" with another element interposed therebetween is also included. Note that the phrase "including" a certain component does not mean that other components are excluded, but means that other components may be included, unless otherwise specified.

Referring to fig. 1, an orthostatic hypotension detection system 100 using a heart rate based machine learning algorithm according to an embodiment of the present invention may include an input 110, a determination 120, and a learning data database 130, and may further include a display 140.

The input 110 is used to receive variables for orthostatic hypotension detection.

According to an embodiment, the input 110 may receive variables of at least one of an E-I difference and an E: I ratio calculated from the age, blood pressure, heart rate of the patient, and a valsalva ratio calculated from valsalva maneuvers, particularly preferably wherein the received variables include the E-I difference, the E: I ratio, and the valsalva ratio. Here, E means exhalation, and I means inhalation. Further, the heart rate of the patient is the heart rate measured during deep breathing, in a supine position at rest, or in a standing position at rest. That is, according to an embodiment of the present invention, it can be determined whether the patient has orthostatic hypotension based on the heart rate that can be acquired by the non-posture stimulation including the supine position and the standing position at rest without the HUT.

For example, the input part 110 may receive age and blood pressure information of the patient corresponding to basic information of the patient from a patient terminal, a medical staff terminal, or an external server (hospital information system). Here, the blood pressure information may include reference systolic and diastolic blood pressure information.

Further, the input 110 may receive an E-I difference and an E: I ratio calculated from heart rates measured during non-posture stimulation, or may also receive heart rate information measured during non-posture stimulation to calculate an E-I difference and an E: I ratio. For this purpose, the input 110 may be provided with processing means for calculating the E-I difference and the E: I ratio.

In particular, when using a heart rate measured during deep breaths according to an embodiment, a heart rate range is measured during deep breaths (e.g., 6 breaths per minute), based on which an E-I difference may be calculated by subtracting a minimum heart rate during expiration from a maximum heart rate during inspiration for each 6 breathing cycle, and an E-I ratio may be calculated by a ratio of a longest R-R interval during expiration to a shortest R-R interval during inspiration.

Further, the input unit 110 may receive a valsalva ratio calculated by a valsalva maneuver, or may receive heart rate information measured according to the valsalva maneuver to calculate the valsalva ratio.

Specifically, in accordance with the valsalva maneuver, the patient is asked to blow through a mouthpiece connected to a pressure gauge to measure the heart rate range while maintaining a pressure of 40mmHg for 15 seconds in a comfortably reclined posture, and the valsalva ratio can be calculated from the value of the maximum R-R interval divided by the minimum R-R interval.

According to an embodiment, the heart rate measurement during deep breathing, at rest or according to the valsalva maneuver can be performed by a wearable measuring device wearable on the patient's body, for which purpose the content guiding the deep breathing or valsalva maneuver is provided by an additionally provided display 140 and the heart rate during the stimulation can be measured. Thus, the heart rate of the patient during deep breathing, at rest or during non-postural stimulation according to valsalva maneuvers can be measured in daily life, and on the basis of this, it can be more conveniently determined whether orthostatic hypotension is present.

The determination section 120 is configured to determine whether the patient has orthostatic hypotension according to a machine learning algorithm 121 based on the variables received through the input section 110.

According to one embodiment, the machine learning algorithm 121 may receive at least one of the patient's age, blood pressure, E-I difference, E: I ratio, and Walsalva ratio, and based thereon, may determine whether orthostatic hypotension is present. In particular, it is preferred that the variables received by the machine learning algorithm 121 include the E-I difference, the E: I ratio, and the Wallsaw ratio.

For this, the machine learning algorithm 121 may perform learning in advance using learning data stored in the pre-established learning data database 130. For example, the machine learning algorithm 121 may learn to determine whether orthostatic hypotension is present using learning data (i.e., age, blood pressure, E-I difference, E: I ratio, and valsalva ratio data for patients diagnosed with OH by HUT and non-OH patient groups) stored in the learning data database 130.

According to an embodiment, the machine learning algorithm 121 may use a machine learning algorithm such as an SVM (support vector machine) algorithm, a KNN (K nearest neighbor) algorithm, and a random forest algorithm, but is not limited thereto and may be used by selecting from a variety of learning algorithms known to those of ordinary skill in the art.

Table 1 shows performance according to the kind of algorithm used in the machine learning algorithm, specifically, classification accuracy in the case where the SVM algorithm, the KNN algorithm, and the random forest algorithm are applied using 5 kinds of input variables including the age, the blood pressure, the E-I difference, the E: I ratio, and the valsalva ratio of the patient, respectively.

[ Table 1]

Algorithm	Precision degree	Recall rate	Accuracy of
				SVX	0.83	0.88	0.84
KNN	0.83	0.88	0.84
				Random forest	0.94	0.88	0.91

As can be seen from table 1, the accuracy is highest when using the random forest algorithm.

As described above, by using the machine learning algorithm 121 for learning, it is possible to accurately determine whether or not orthostatic hypotension is present in time in daily life on the basis of the variables received in real time through the input unit 110.

Referring to fig. 2, an orthostatic hypotension detection system 200 using a heart rate based machine learning algorithm according to an embodiment of the present invention may include a wearable measurement device 210, a processing apparatus 220, a determination part 230, and a learning data database 240, and may further include a display part 250.

The wearable measurement device 210 may be worn on the patient's body and measure the patient's heart rate.

The processing means 220 may calculate the E-I difference and the E: I ratio from the heart rate measured by the wearable measurement device 210, and may calculate the valsalva ratio from the heart rate measured by the wearable measurement device 210 according to a valsalva maneuver.

The determination section 230 may determine whether the patient has orthostatic hypotension according to a learned machine learning algorithm based on the E-I difference, the E: I ratio, and the valsalva ratio calculated by the processing device 220.

The display device 250 may provide content that guides deep breathing or valsalva maneuvers.

Since the specific functions of each component of the orthostatic hypotension detection system 200 using the heart rate-based machine learning algorithm shown in fig. 2 are the same as those described above with reference to fig. 1, redundant description thereof will be omitted.

Referring to fig. 3, first, age and blood pressure information of a patient may be acquired (step S31), an E-I difference and an E: I ratio are calculated from the measured heart rate (step S32), and a valsalva ratio is calculated according to valsalva maneuvers (step S33). Here, the heart rate of the patient is the heart rate measured during deep breathing, measured in a supine position at rest, or measured in a standing position at rest.

In fig. 3, steps sequentially performed in the order of steps S31 to S33 are shown, but this is merely illustrative and the order is not limited thereto, and it is sufficient to acquire the age, blood pressure, E-I difference, E: I ratio, and valsalva ratio of the patient.

Thereafter, it may be determined whether or not orthostatic hypotension is suffered according to a machine learning algorithm based on the information acquired in steps S31 to S33, i.e., the age, blood pressure, E-I difference, E: I ratio, and valsalva ratio of the patient (step S34).

Since a specific method of performing each step of the orthostatic hypotension detection method using the heart rate-based machine learning algorithm shown in fig. 3 is the same as that described above with reference to fig. 1, a redundant description thereof will be omitted.

Further, the orthostatic hypotension detection method using a machine learning algorithm based on heart rate shown in fig. 3 may be performed by a processing device capable of performing a machine learning algorithm.

According to an embodiment of the present invention, there may be provided a computer-readable storage medium having recorded thereon instructions executable by a processor to perform each step of the orthostatic hypotension detection method using a heart rate based machine learning algorithm shown in fig. 3.

According to another embodiment of the present invention, the orthostatic hypotension detection method using a heart rate-based machine learning algorithm shown in fig. 3 may be implemented in the form of software that may be carried on a wearable measurement device worn on the body of a patient and measuring the heart rate of the patient. Thus, the heart rate of a patient may be measured in daily life by a wearable measuring device, on the basis of which it may be used as an OH screening tool for determining whether or not orthostatic hypotension is present.

The present invention is not limited to the above-described embodiments and drawings. It is apparent to those having ordinary knowledge in the technical field to which the present invention pertains that substitutions, variations, and modifications of the constituent elements according to the present invention can be made within a scope not exceeding the technical spirit of the present invention.

Claims

1. An orthostatic hypotension detection system using a heart rate based machine learning algorithm, comprising:

an input part for receiving variables including at least one of an E-I difference and an E: I ratio calculated from the age, blood pressure and heart rate of a patient and a Wal-Save ratio calculated from the Wal-Save movement; and a process for the preparation of a coating,

a determination section that determines whether the patient has orthostatic hypotension according to a learned machine learning algorithm based on the variables received through the input section.

2. The orthostatic hypotension detection system using heart rate based machine learning algorithm of claim 1,

the heart rate is measured during a patient's deep breath, in a supine position at rest, or in a standing position at rest.

3. The orthostatic hypotension detection system using heart rate based machine learning algorithms of claim 1, further comprising:

and a display unit for providing contents for guiding deep breathing or valsalva movements.

4. An orthostatic hypotension detection system using a heart rate based machine learning algorithm, comprising:

a wearable measurement device worn on a patient and measuring a heart rate of the patient;

a processing device for calculating an E-I difference and an E: I ratio from the heart rate measured by the wearable measuring device, and calculating a valsalva ratio from the heart rate measured by the wearable measuring device according to a valsalva maneuver; and a process for the preparation of a coating,

a determination unit for determining whether the patient has orthostatic hypotension based on the E-I difference, the E: I ratio and the Walsajou ratio calculated by the processing unit according to a learned machine learning algorithm.

5. The orthostatic hypotension detection system using heart rate based machine learning algorithms of claim 4, further comprising:

a display device providing content for guiding a deep breath or a valsalva maneuver.

6. A method of orthostatic hypotension detection using a heart rate based machine learning algorithm, comprising the steps of:

calculating an E-I difference and an E: I ratio from the heart rate of the patient;

calculating a valsalva ratio according to the valsalva maneuver; and a process for the preparation of a coating,

based on the E-I difference, the E: I ratio, and the Walsakava ratio, a machine learning algorithm determines whether orthostatic hypotension is present.

7. The orthostatic hypotension detection method using heart rate based machine learning algorithm of claim 6, further comprising the steps of:

the age and blood pressure information of the patient is acquired,

wherein the determining whether to have orthostatic hypotension further determines whether to have orthostatic hypotension taking into account age and blood pressure information of the patient.

8. A wearable measurement device to be worn on a patient's body and to measure the patient's heart rate,

software implementing the orthostatic hypotension detection method using a heart rate based machine learning algorithm of claim 6 is loaded, and it is determined by the software whether the patient has orthostatic hypotension based on the measured heart rate.